Solar Energy Vol. 62, No. 6, pp.

419–427, 1998
© 1998 Elsevier Science Ltd
PII: S0038-092X(98)00027-9 All rights reserved. Printed in Great Britain
0038-092X/98 $19.00+0.00

SOLAR POTENTIAL FOR WATER HEATING EXPLORED BY GIS

D. VOIVONTAS,* G. TSILIGIRIDIS** and D. ASSIMACOPOULOS*

* National Technical University of Athens, Section II: Process Analysis and Plant Design,
9 Heroon Polytechniou St., Zografou Campus, GR-157 80 Athens, Greece
** Aristotle University of Thessaloniki, Thessaloniki, Greece

Received 21 May 1997; revised version accepted 4 February 1998

Communicated by ARI RABL

Abstract—A method is presented for the estimation of solar energy and market potential for water heating
in the residential sector. The model was developed under a Geographical Information System and provides
the tools to handle the spatial and time discrepancies of solar radiation and energy demand. A geographic
database with climatic data is used for estimating efficiencies and monthly/annual coverage of water
heating load. Financial analysis is conducted on the basis of the energy production cost and the Net
Present Value of the investment. Different financial scenarios are considered and the expected energy
yields from a large-scale deployment of solar thermal systems in the residential sector of Greece are
assessed. © 1998 Elsevier Science Ltd. All rights reserved.

1. INTRODUCTION this task, Geographic Information Systems

The estimation of solar potential and the evalua-
tion of energy demand for specific end-use
activities are the basic steps to determine the
best policies which can facilitate the large-scale
deployment of a wide array of solar energy
applications.
Solar radiation is highly dependent on site
location and regional climatological conditions.
The mathematical framework for calculating
hourly, daily and monthly average solar radia-
tion as well as the basic design principles for
solar collectors and energy conversion technol-
ogies are presented by Magal (1990), Duffie
and Beckman (1980) and Kreider (1979).
Recent research efforts that focus on methods
which model and estimate spatial and temporal
distribution of solar radiation can be found in
Barr et al. (1996), Islam and Exell (1996),
Skarveit et al. (1996), Olseth et al. (1995) and
Lalas et al. (1982). Research on system design
and models assessing the performance of specific
solar systems are presented, among others, by
Panteliou et al. (1996), Ghaddar et al. (1997),
Tsilingiris (1996) and Michaelides and Wilson
(1997).
The relation of specific geographical charac-
teristics, technological and financial factors to
local energy market potential, however, has not
yet been addressed. The potential contribution
of solar energy to the energy demand of specific
end-use activities and the potential market for
solar applications have to be analysed and the
expected profits from a large-scale deployment
of solar energy systems have to be identified. In
419
(GIS) technology is a new promising approach
which can reveal the spatial and time discrepan-
cies of solar potential and energy demand.
Solar Domestic Hot Water Systems
(SDHWS ) is the most simple and well devel-
oped solar energy application. A large potential
market exists, since, even in regions with favour-
able sunshine conditions the penetration of
SDHWS is rather limited ( EUREC Agency,
1996). In the present work, the matching of
water heating energy demand to solar potential
is investigated within a GIS environment. Well
known methods are integrated into a consistent
GIS model: the estimation of the water heating
energy demand is based on demographic data
attributed to geographic regions; the calculation
of the regional solar radiation is based on the
methods presented by Magal (1990) and Duffie
and Beckman (1980) and makes use of inter-
polated meteorological data following the tech-
niques presented by Barr et al. (1996); the
assessment of the expected energy savings from
the integration of SDHWS into the regional
market is estimated by the f-chart method and
demographic data. Finally, the expected finan-
cial profits are estimated on the basis of
common economic indices. Results from a large-
scale deployment of SDHWS in the domestic
sector in Greece are presented as a case study.
In Section 2 the GIS approach of determining
solar potential is presented, in Section 3 the
obtained results are given, and in Section 4
discussion of results and methodology is
presented.
420 D. Voivontas et al.

2. METHODOLOGY solar radiation at the top of the atmosphere

and estimations of the clearness index from
GIS environments handle geographical enti-
sunshine measurements for each residential
ties, attributes defining their characteristics and
area. The inputs and results of this step are
methods determining their behaviour and inter-
formed as attributes of geographical objects
action. All possible geographical entities can be
(points) corresponding to residential areas. GIS
modelled within a GIS environment as point,
provides the built-in method to estimate the
line or polygon objects. Built-in attributes and
latitude of each residential area while additional
methods are used to create, store and retrieve
methods have been developed for the estimation
geographic objects. Attributes are stored in the
of the sunshine duration. The methods that
form of database records and are referenced
handle the calculations in this step make use of
to the corresponding geographical objects.
the attributes of the residential areas’ objects.
Methods can be categorised to those that handle
Screening effects from the surrounding horizon
geographic operations and set a minimum con-
have not been introduced for the estimation of
nectivity between different sets of objects and
total radiation reaching a tilted surface because
those that handle database calculations and
topography effects (i.e. mountains, hills, large
expand the interactions of objects. The user has
obstacles blocking the sun) become substantial
the flexibility to create different sets of objects
on latitudes more than 60° and because other
defined on the basis of common characteristics
effects (e.g. trees or buildings near the collector,
and models for extensive calculations using the
roof orientation) would be useful in a roof-by-
built-in framework of object behaviours.
roof analysis which is not pursued in the present
The model for the estimation of solar energy
work. Methods to introduce such effects on
potential and energy demand for water heating
solar radiation estimations can be found else-
is built in a GIS environment with the objective
where (Olseth et al., 1995).
to exploit its powerful features in handling all
Reliable long-term sunshine measurements
the parameters with geographic variability that
are sparse and available only for regions where
influence solar radiation ( latitude, sunshine
meteorological stations are operating. In the
duration, clearness index) and energy demand
regions where there are no data available the
(population, number of households). An addi-
sunshine duration has been estimated using a
tional objective is to reveal the spatial interrela-
simple weighted interpolation method (inverse
tions between solar radiation availability and
squared distance) of measured values from
the expected energy savings achieved from a
meteorological stations with similar climatologi-
large-scale implementation of SDHWS. The
cal conditions, as suggested by Barr et al.
GIS that has been used is MAPINFO
(1996). The ratio of sunshine duration to the
Professional for Windows.
maximum sunshine duration for each residential
The model structure for the estimation of
area was used to obtain the clearness index for
solar potential for water heating is presented in
the area. The clearness index, defined as the
Fig. 1. The evaluation procedure consists of
proportion of the radiation at the top of the
three steps: the estimation of solar radiation,
atmosphere which reaches the surface of the
the estimation of energy demand for water
earth, was then used to estimate the average
heating, and the financial analysis. The appro-
monthly radiation on horizontal and tilted sur-
priate database structure has been dictated by
faces. Additional attributes for the residential
the inputs needed by the models and the
areas’ objects have been used to host the esti-
obtained results at each step of the methodol-
mated values of clearness index and average
ogy. In the following paragraphs, the calcula-
monthly solar radiation and the methods to
tions in each step of the evaluation procedure
calculate the solar radiation have been devel-
are outlined and the geographic objects, attri-
oped in the GIS environment.
butes and methods used are described.
The meteorological stations used for the esti-
mation of clearness index are selected on the
2.1. Estimation of solar radiation basis of similar climatological conditions. A
Among the most important factors influenc- digital map is developed, setting the boundaries
ing the solar radiation reaching a horizontal or of regions of the same sunshine conditions. The
tilted surface in a region is the latitude and the regions are a subclass of polygon objects with
sunshine duration. Solar radiation on horizontal boundaries defined by the large mountain chains
or tilted surfaces is evaluated by the calculated (considered to have a significant impact on the
 Solar potential for water heating explored by GIS
Fig. 1. Flowchart for the evaluation of the solar potential for water heating in the domestic sector.
cloudiness of surrounding regions). The meteo-
rological stations are a subclass of point objects
and their attributes contain the measured values
for sunshine duration. For each residential area
within a region, the clearness index is estimated
using data from the meteorological stations
sited in the same region. The identification of
the residential areas and meteorological stations
within the same region and the calculation of
the distance between meteorological stations
and residential areas is performed by geographic
operations that provide the relative geographi-
cal location of point and region objects. The
methods for the implementation of the inverse
square distance interpolation of sunshine dura-
tion measurements and the estimation of the
clearness index have been added to the residen-
tial areas’ objects.
2.2. Energy demand for water heating
All households residing in one and two-storey
buildings are considered as potential users of
solar thermal systems for hot water production.
The solar radiation that can be harvested is
restricted by space or market availability (area
available on the roofs of buildings and number
of households which have already installed
SDHWS).
421
The average monthly energy needs for water
heating of a household (Appendix, eqn (1)) are
estimated taking into account the distribution
of household size (number of household mem-
bers) and the assumption that energy needs for
water heating are proportional to the household
members. The monthly energy demand is a
function of the temperature in the water distri-
bution networks. The results are stored in a
database which is attached to the geographical
objects corresponding to the residential areas.
The monthly average percentage of energy
needs for water heating covered by the solar
system is calculated by the f-chart method
(Beckman et al., 1977). Inputs for the f-chart
method are the average monthly solar radiation
on a tilted surface, the monthly energy needs
for water heating, the technical characteristics
of the solar thermal system and the average
monthly ambient temperature. The f-chart
method calculation is performed on the attri-
butes of the residential areas’ objects and identi-
fies the spatial interrelation of energy demand
and solar energy potential. The dependence of
energy needs covered by a SDHWS on ambient
temperature is addressed taking also into
account the dependence of ambient temperature
on altitude. It is assumed that the ambient
422 D. Voivontas et al.
temperature decreases with the altitude at a rate
of 1°C/103 m (Seinfeld, 1986). The altitude of
each residential area has been set equal to that
of the nearest iso-altitude line. Monthly ambient
temperature data are sparse and a weighted
inverse squared distance method is used to
interpolate measured values. Spatial and tempo-
ral variability of solar radiation, ambient tem-
perature, water temperature and water heating
demand for each type of households have to be
introduced by the GIS in order to correctly
apply the f-chart method in each residential
area and type of household.
2.3. Financial analysis
The expected energy savings as well as the
thermal energy production cost are calculated
in the financial analysis step. The GIS is used
to address the geographic allocation of consum-
ers and energy needs in conjunction with the
solar radiation variability. In addition, GIS is
used to present the geographic variability of
thermal energy cost which mainly depends on
the energy needs covered by the SDHWS.
Fig. 2. Present penetration of Solar Thermal Systems for Water Heating.
The expected energy savings represent the
amount of thermal energy produced by electric-
ity that can be replaced by solar energy and are
estimated with the assumption that all house-
holds in one or two-storey buildings, without a
SDHWS, will install a solar thermal system.
The expected total energy savings depend on
the geographic distribution of households in
areas of different solar radiation and ambient
temperature.
The annual cost of thermal energy produced
by a SDHWS is estimated taking into account
the amortisation of the investment cost and
depends on the percentage of coverage of water
heating load. The Net Present Value (NPV ) for
various discount rates is estimated considering
that the cost of a SDHWS represents the capital
cost of the investment and the amount of
electricity savings represents an annual income
which otherwise would be spent for the hot
water production needs. In addition, sensitivity
analysis for the capital cost and the efficiency
of the SDHWS is conducted for a specific
investment.
 Solar potential for water heating explored by GIS 423

3. CASE STUDY the current penetration of solar thermal systems

The financial incentives for SDHWS offered
in Greece during the last few decades boosted
the market in the domestic sector. In 1995 the
total installed collector area, in Greece, was
about 2,000,000 m2 and 95% of them are
SDHWS. More than 600,000 households use a
SDHWS which provides about 80% of the
energy needed for hot water and thus reduces
CO emissions by 1,500,000 ton each year
2 data on the existing penetration of solar ther-
( European Commission, 1996). The level of the
existing penetration is in the range of 2–30% of
households and varies from region to region.
This is considered very low compared to regions
with similar climate (Cyprus, Israel ) where a
penetration of more than 90% has been realised.
A wide market survey performed by the Greek
Solar Industries Association has found that
more than 90% of the owners of solar systems
are satisfied and if they replace the old solar
system they will invest in a solar system again.
Fig. 2 presents the geographic distribution of
for water heating in Greece.
The developed methodology has been used
for the estimation of the solar potential for
water heating in the residential sector in Greece.
The data used are:
A digital map with the location of residential
areas and census data on the number of house-
holds and buildings up to two storeys.
A digital map of prefecture boundaries with
mal systems.
A digital map of altitude lines.
Radiation measurements on horizontal
surface.
A digital map of the meteorological stations’
locations with measurements of sunshine dura-
tion, and ambient temperature.
Technical characteristics of typical solar ther-
mal systems.
Monthly average water temperature in the
water distribution networks.

Fig. 3. Annual solar radiation on 40° tilted surface.

424 D. Voivontas et al.
Table 1. Technical characteristics of a typical SDHWS using
a flat collector with selective surface and one glass
Fr×U 4.07 ( W m−2 °C−1)
L
Fr×(ta) 0.69
n
(Fr∞/Fr) 0.95
(ta)/(ta) 0.94
n 75–150 l for families with 1–6 members (for the
Collector tilt 40°
The location of all the municipalities in
Greece is recorded on a layer (map) and each
geographic object is indexed to the database
record with all data for the corresponding resi-
dential area. Solar radiation on a horizontal or
tilted surface has been estimated for all residen-
tial areas of the country. Fig. 3 presents the
monthly average solar radiation on a 40° tilted
surface.
The monthly coverage of water heating
demand of all households with 1–6 members
living in one or two-storey buildings and not
already possessing a SDHWS (75% of the total
number of households in Greece) has been
estimated. The technical characteristics of a
typical flat-plate, single-glazed selective solar
Fig. 4. Mean annual coverage of water-heated load.
collector, which has been used in the present
application, are presented in Table 1. The area
of the collector covering the hot water needs of
a family is estimated at 1 m2 per household
member and the tank volume ranges from
Greek climatic conditions). Fig. 4 presents the
mean annual coverage of water heating load for
each community in Greece.
The investment cost for a SDHWS has been
estimated using a correlation derived from
market prices of various systems available in
Greece (Appendix, eqn (2)). The investment
cost includes the collector, hot water tank and
installation costs. The cost of the thermal energy
produced by the SDHWS has been estimated
for each city or village and each type of house-
hold considering that the lifetime of the invest-
ment is 15 years, the discount rate is 10% and
the inflation rate is 8%. The total energy savings
have been estimated for each city using the total
number of households. Table 2 presents the
energy savings by solar thermal systems for
each category of households distributed accord-
 Solar potential for water heating explored by GIS 425

Table 2. Energy savings (GWh yr−1) using solar thermal systems

th
Energy cost Number of members per household Total
(mECU kW h−1) 2 3 4 5 6 (GWh yr−1)
th
36–40 0 0 0 10 117 127
40–44 0 0 7 339 152 127
44–48 0 0 668 225 16 909
48–52 0 45 461 1 0 507
52–56 0 465 64 0 0 529
56–60 0 269 0 0 0 269
60–64 27 13 0 0 0 40
64–68 346 0 0 0 0 346
68–72 110 0 0 0 0 110
72–76 193 0 0 0 0 193
76–80 12 0 0 0 0 12
Total 690 795 1204 580 291 3169

ing to the thermal energy production cost of a
SDHWS. Fig. 5 presents the geographical
distribution of energy cost for three-member
households.
A comparison between the cost of thermal
energy produced by an electrical system and a
SDHWS for each category of households is
summarised in Table 3. The electricity cost for
individual consumers throughout Greece is
0.07 ECU kW h−1. The cost of the thermal
energy produced by an electrical system is esti-
mated considering the amortisation of the initial
investment and the energy needs for each house-
hold type.
The effect of capital cost and efficiency loss
of the system on the viability of the investment

Fig. 5. Cost of thermal energy produced by SDHWS for three-member families. The electricity cost for the same case is
82.5 mECU kW h−1.
426 D. Voivontas et al.

Table 3. Cost of thermal energy produced by electricity and this case, the energy produced by the SDHWS
solar domestic hot water systems (mECU kW h−1)
th is assumed to be reduced after the fourth year
Members Thermal energy Thermal energy of operation with an annual rate of 5–10%, a
per cost with cost with fact which results in a diminution of the annual
family electricity SDHWS
cash flow.
2 85.3 60–80 The obtained results show that the annual
3 82.5 48–64 demand of energy for water heating for a typical
4 81.1 40–56 family can be covered by solar thermal systems
5 80.2 36–52
6 79.7 36–48 at a percentage of 65–87% (varying by type of
household and region).
The energy cost is higher for two-member
is demonstrated by the sensitivity analysis for families and decreases as the number of mem-
those parameters. Fig. 6 presents the sensitivity bers per family is increased while the percentage
analysis for the capital cost for a specific of water heating demand covered by solar sys-
SDHWS investment for a four-member family tems is increased.
in Athens. It is assumed that the capital cost is The current energy savings using of
reduced by 10 and 20%. Fig. 7 presents the 2×106 m2 of solar collectors are about
sensitivity analysis for the efficiency of the 1200 GWh yr−1 of thermal energy. The present
system for a specific installation in Athens. In analysis shows that additional 3169 GWh yr−1

Fig. 6. Sensitivity analysis for the capital cost for a four-member family in Athens.

Fig. 7. Sensitivity analysis for the efficiency of SDHWS for a four-member family in Athens. It is considered that the
efficiency of the system is steadily reduced after the fifth year.
 Solar potential for water heating explored by GIS
of thermal energy savings could be obtained if
the 75% of remaining households install a
SDHWS.
The cost of thermal energy produced by
SDHWS is well below the cost of electricity for
all types of households and their installation is
a good investment for almost all the regions of
Greece. Financial incentives with the aim to
reduce the capital cost of SDHWS would make
such installations even more attractive to
consumers.
4. CONCLUSIONS
The GIS method for the estimation of solar
potential for water heating, presented in this
work, takes into account the geographic distri-
bution of solar radiation and the market for
solar thermal systems and identifies the alloca-
tion of the expected energy savings and profits
from a large-scale deployment of SDHWS. The
adoption of the GIS environment provides some
new spatial insight on the results obtained by
well known methods applicable to each stage
of the analysis. The matching of solar radiation
and energy demand on a geographic basis offers
the necessary information for energy policy
decisions.
The spatial variability of the energy demand
covered by a SDHWS revealed by the introduc-
tion of the geographic discrepancies of all spa-
tially-dependent parameters makes the results
of the proposed methodology even more interes-
ting to policy makers.
Acknowledgements—Partial financial support from the
Economic Potential Use of Renewable Energies ( EPURE),
EEC, DG XII and Program APAS CT94-0054, 1995–96 is
gratefully acknowledged.
REFERENCES
Barr A. G., McGinn S. M., Cheng S. B. (1996) A compari-
son of methods to estimate daily global solar irradiation
from other climatic variables on the Canadian prairies.
Solar Energy 56, 3, 213–224.
Beckman W. A., Klein S. A. and Duffie J. A. (1977)
Solar Heating Design by the f-chart Method. Wiley
Interscience, New York.
427
Duffie J. A. and Beckman W. A. (1980) Solar Engineering
of Thermal Processes. John Wiley and Sons, New York.
EUREC Agency (1996) The Future for Renewable Energy,
Prospects and Directions. James and James, London.
European Commission, European Solar Industry Federa-
tion (1996) Sun in Action. ALTENER Programme.
Ghaddar N. K., Shihad M. and Bdeir F. (1997) Modelling
and simulation of solar absorption system performance
in Beirut. Renewable Energy 10, 4, 539–558.
Islam M. R., Exell R. H. (1996) Solar radiation mapping
from satellite image using a low cost system. Solar
Energy 56, 3, 225–237.
Kreider J. (1979) Medium and High Temperature Solar Pro-
cesses. Academic Press, New York.
Lalas D., Pissimanis D. and Notaridou V. (1982) Methods
of estimation of the intensity of solar radiation on a
tilted surface and tabulation data for 30, 40 and 60
degrees in Greece. Dept. of Meteorology, University
of Athens.
Magal B. S. (1990) Solar Power Engineering. McGraw–Hill,
New Delhi.
Michaelides I. M., Wilson D. R. (1997) Simulation studies
of the position of the auxiliary heater in termosyohon
solar water heating systems. Renewable Energy 10, 1,
35–42.
Olseth J. A., Skarveit A. and Zou H. (1995) Spatially con-
tinuous mapping of solar resources in a complex high
latitude topography. Solar Energy 55, 6, 475–485.
Panteliou S., Dentsoras A. and Daskalopoulos E. (1996)
Use of expert systems for the selection and the design
of solar domestic hot water systems. Solar Energy 57,
1, 1–8.
Seinfeld J. H. (1986) Atmospheric Chemistry and Physics of
Air Pollution. John Wiley and Sons, New York.
Skarveit A., Olseth J. A., Czeplak G. and Rommel M.
(1996) On the estimation of atmospheric radiation from
surface meteorological data. Solar Energy 56, 4,
349–359.
Tsilingiris P. T., Design and performance of large scale low
cost solar water heating system. Renewable Energy 9, 1,
2, 3, 4, (1996) 617–621.
APPENDIX
Energy needs. The average monthly energy for water heat-
ing needed for each household is estimated by eqn (1):
L =N ×N ×V ×r×C ×(T −T ) (1)
w p d w p w m
Where L =energy load for water heating, N =number of
w p
persons per household, N =number of days per month,
d
V =daily average hot water consumption for each person
w
(40 l ), r=water density, C =water specific heat,
p
T =required hot water temperature (55°C ) and
w
T =average monthly temperature of water in the distribu-
m
tion network.
Financial analysis The capital cost of the investment
in ECU is estimated by eqn (2):
C =177×A +333 (2)
c c
Where C =capital cost of the solar thermal system and
c
A =collector surface.
c